With the development of new generation scalable video compression standard and continuous improvement of the calculation capability and the storage capability of wireless equipment, applications based on wireless network video transmission are increasing. However, characteristics of a wireless channel, such as a time-variant characteristic, a high bit error rate, a low bandwidth and the like, seriously restrict and affect the quality of video received by a receiving end. As a multi-input-multi-output (MIMO) technology is put forward, the difficult problem is alleviated to a great extent.
Video stream with a multilayer structure generated by using a scalable video coding (SVC) technology is very suitable to be transmitted on an MIMO system. Different layers of the video stream are scheduled to different sub-channels and transmitted simultaneously, so as to support a high video bit rate. However, the time-variant characteristic of the wireless channel determines that at different moments, the quality of each sub-channel is not completely same and further has the time-variant characteristic. Moreover, the importance of each layer of data in an SVC video stream is not completely same either, for example, data of a base layer are most important; data of a first enhancement layer come second, and so on.
In the process of implementing technical solutions of embodiments of this application, inventors of this application at least discover that the prior art has the following technical problems:                if each video layer is equally treated indiscriminately according to the above-mentioned method, when the data of the base layer are scheduled to a sub-channel with poor quality condition, the quality of video received by the receiving end can be seriously affected.        
Moreover, in an MIMO broadcast channel, each transmitting antenna probably corresponds to multiple sub-channels. Accordingly, the signal to noise ratio or packet loss rate of each antenna is actually determined by a vector, rather than a traditional single scalar. The vector can not be compared in size or sequence, so that the quality of the antenna can not be measured, and finally, an optimal video-antenna matching relation is difficult to be sought. Therefore, a video-antenna matching relation is generally randomly selected in the prior art, but the transmission quality can not be ensured in this case.